Advancing Automated In-Isolation Validation in Repository-Level Code Translation

2025/11/26 by Ke, Kaiyao, Ibrahimzada, Ali Reza, Pan, Rangeet +2 · 1 citation
Computer Science · Decision Sciences · #FOS: Computer and information sciences #Programming Languages (cs.PL) #Scientific Computing and Data Management #Software Engineering (cs.SE) #Software Engineering Research #Software Testing and Debugging Techniques

paper · doi:10.48550/arxiv.2511.21878

openalex publication_date 2025/11/26 · openalex created_date 2025/12/03 · openalex updated_date 2026/07/28

Abstract

Repository-level code translation aims to migrate entire repositories across programming languages while preserving functionality automatically. Despite advancements in repository-level code translation, validating the translations remains challenging. This paper proposes TRAM, which combines context-aware type resolution with mock-based in-isolation validation to achieve high-quality translations between programming languages. Prior to translation, TRAM retrieves API documentation and contextual code information for each variable type in the source language. It then prompts a large language model (LLM) with retrieved contextual information to resolve type mappings across languages with precise semantic interpretations. Using the automatically constructed type mapping, TRAM employs a custom serialization/deserialization workflow that automatically constructs equivalent mock objects in the target language. This enables each method fragment to be validated in isolation, without the high cost of using agents for translation validation, or the heavy manual effort required by existing approaches that rely on language interoperability. TRAM demonstrates state-of-the-art performance in Java-to-Python translation, underscoring the effectiveness of its integration of RAG-based type resolution with reliable in-isolation validation.

Citations

• Translating Large-Scale C Repositories to Idiomatic Rust
• EffiReasonTrans: RL-Optimized Reasoning for Code Translation
• RustAssure: Differential Symbolic Testing for LLM-Transpiled C-to-Rust Code
• Adversarial Agent Collaboration for C to Rust Translation
• MatchFixAgent: Language-Agnostic Autonomous Repository-Level Code Translation Validation and Repair
• Integrating Rules and Semantics for LLM-Based C-to-Rust Translation
• EvoC2Rust: A Skeleton-guided Framework for Project-Level C-to-Rust Translation
• TypyBench: Evaluating LLM Type Inference for Untyped Python Repositories
• Search-Based Multi-Trajectory Refinement for Safe C-to-Rust Translation with Large Language Models
• CRUST-Bench: A Comprehensive Benchmark for C-to-safe-Rust Transpilation
• Program Skeletons for Automated Program Translation
• RustMap: Towards Project-Scale C-to-Rust Migration via Program Analysis and LLM
• C2SaferRust: Transforming C Projects into Safer Rust with NeuroSymbolic Techniques
• Syzygy: Dual Code-Test C to (safe) Rust Translation using LLMs and Dynamic Analysis
• AlphaTrans: A Neuro-Symbolic Compositional Approach for Repository-Level Code Translation and Validation
• On Mitigating Code LLM Hallucinations with API Documentation
• Rectifier: Code Translation with Corrector via LLMs
• VERT: Verified Equivalent Rust Transpilation with Large Language Models as Few-Shot Learners
• Exploring and Unleashing the Power of Large Language Models in Automated Code Translation
• The Emergence of Large Language Models in Static Analysis: A First Look through Micro-Benchmarks
• DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence
• CodeTransOcean: A Comprehensive Multilingual Benchmark for Code Translation
• On the Evaluation of Neural Code Translation: Taxonomy and Benchmark
• Extending Context Window of Large Language Models via Positional Interpolation
• StructCoder: Structure-Aware Transformer for Code Generation
• Leveraging Automated Unit Tests for Unsupervised Code Translation
• Unsupervised Translation of Programming Languages
• Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
• Tree-to-tree Neural Networks for Program Translation
• SpecTra: Enhancing the Code Translation Ability of Language Models by Generating Multi-Modal Specifications

Cited by

• Code Quality Analysis of Translations from C to Rust
• MatchFixAgent: Language-Agnostic Autonomous Repository-Level Code Translation Validation and Repair

Related

• MatchFixAgent: Language-Agnostic Autonomous Repository-Level Code Translation Validation and Repair
• Scalable, Validated Code Translation of Entire Projects using Large Language Models
• ClassEval-T: Evaluating Large Language Models in Class-Level Code Translation
• Program Skeletons for Automated Program Translation
• AVATAR: A Parallel Corpus for Java-Python Program Translation